process-reviews

Warn

Audited by Snyk on Mar 4, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.80). The skill explicitly ingests reviewer PDFs (auto-discovered from paths like to-sort/*.pdf or correspondence/referee-reviews/...) and the Phase 2 instruction "Read the reviews PDF using the Read tool" requires the agent to parse untrusted, user-generated reviewer text which is then used to drive analysis, prioritization, and strategy decisions, so third‑party content can materially influence behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 4, 2026, 07:18 PM